Data exploration is a process of discovering generalised rules and knowledge contained in databases, based on statistical methods and artificial intelligence techniques. The idea of data exploration consists in using the computer’s speed to find regularities in accumulated data that are hidden from man (due to time constraints). There are many data exploration techniques derived from well-established fields of science such as statistics and machine learning. One of the most important issues pertaining to the field of machine learning is the selection of a classification method.

How can I use data exploration in e-commerce marketing?

To determine the target Customers who are very likely to make a purchase during the sales campaign, you can use the classification method referred to as a decision tree. Decision trees are a graphic method of supporting the decision-making process. The algorithm of decision trees is also applied in machine learning for acquiring knowledge based on examples. The concept of Bugging consists in building experts for a subset of tasks. In this case, a subset of problems is randomly selected with replacement among all problems to be solved and then an expert is sought for this subset. In this algorithm, a subset is randomly selected from among all elements of the data set, and a prediction model is built for this subset. Next, another subset of vectors is again randomly selected with replacement and another model is built for it. The whole procedure is repeated a few times, and all models that have been built are used for voting at the end.

Decision tree methods include a few algorithms. Our case uses the algorithm named Random Forest, consisting in creating many decision trees based on a random data set. The idea behind this algorithm is to build a group of experts from randomly selected decision trees, where – unlike traditional decision trees – random trees are built following a principle according to which the subset of features analysed in a node is selected at random.

The characteristics of Random Forest Algorithm

Compared to other algorithms, it’s best in terms of accuracy; it works effectively with large databases; it retains its accuracy if there’s no data or it gives an estimate of which variables are important to the classification; there’s no need to prune the trees; forests can be saved and used in the future for another data set; it doesn’t require knowledge of an expert. Single classifiers of the random forest are decision trees.  Random Forest Algorithm is very suitable for analysing a sample where the observation vector is large.

An analysis of customers’ transactions in the online channel was conducted using Random Forest. The algorithm was launched based on the following parameters:

• Purchase made in a city
• Size of the city population
• Number of days since first purchase
• Number of Customer’s orders
• Number of products in order
• Average discount in order

Algorithm results: 886 customers were classified who are very likely to make a purchase during the next sales action.

The post Application of decision trees in sales campaign appeared first on Upsaily blog.

Are you wondering why your Internet services supplier asks you to rate its customer service portal? Are you irritated by e-mail and text messages, or pop-up windows with requests to rate applications’ new functions or services? “Rate it”, “describe it”, and “say if you like it” messages pop out of almost every window of your browser.

Welcome to the era of customers creating products!

Yes, it can become a nuisance, and more so because the wave of products fuelled by customers’ feedback is gaining momentum. How was it possible that before products were developed without your involvement? It was largely for the same reason why so many of them didn’t succeed and you never heard of them. They failed to respond to actual business needs. In the age of easy access to information, analytical systems, and ways to enable quick contact between the maker and the user, companies that develop solutions for you don’t want to join the crowd of losers before the race even starts.

Before, what mattered at the execution phase of a project was three things – time, budget, and scope. Come to think of it, some people would add quality to this mix. First, products were developed A to Z and only when they were ready did they make their way to the market. Then their authors just waited for the results. Results meaning your reaction, dear User. If a product sold well, it meant that it appealed to the buyer, thus becoming successful. Clearly, it’s easy to imagine a scenario where there’s no success, and all effort, investment, and big plans go to the garbage bin, which is why nowadays – in the Time of Agile Product Development – product makers prefer to make sure early enough that what they’re producing will find buyers.

Who might want to know my opinion?

The project or company’s size doesn’t matter – the development of every product requires feedback. This rule is applicable not only to the customer-producer relation, as feedback comes to developers from all directions – users, other teams in the company, and the market – through trends and competition analysis. Feedback from customers is particularly important to companies developing solutions for a very specific group of users or B2B solutions, or companies trying to launch an innovative product which no existing customer has had an opportunity to use. Consequently, developing another chat-type communication system where it’s easier to start the work based on a benchmark of existing solutions and omnipresent opinions about available applications’ pros and cons is very different from creating a product that has no counterpart on the market (the case of Upsaily, which applies customer intelligence and machine learning for data analysis used in marketing). In the case of the latter, companies will try to acquire as much information as they can about your needs, what’s important to you, and how you have dealt with a given issue so far. Based on the information you provide, the development team will be able to create strategies and tools that can not only streamline your workflow but also help you achieve results that are unattainable with your current work methodology.

How can software developers find out about whether a tool appeals to me?

You must know many such methods – from the already mentioned pop-up messages, surveys, and thumbs up and down to text messages asking your opinion. There are many ways to reach the user, depending on the extent of product development. If the product is at early production stages, its authors will need you, the Customer, to provide a bigger chunk of information – in such cases, the most frequently used strategies are meetings and talks in which the customer presents their needs, reservations, and current strategies for dealing with issues. Based on them, a product development roadmap is built, comprising potentially helpful functions. It’s advisable to accept invitations to such meetings, as it often happens that in return for active participation in product development the developers offer a free-of-charge trial version of the system which your competition hasn’t yet seen or an opportunity to create elements tailored to your business.

If the system is already functional, the pool of strategies for acquiring information is expanded with analytical tools that make it possible to check whether and how you use a given solution, how often you visit a specific part of the application or system, and which, if any, parts of the system you look at for a longer time. Such tools enable regular checks and adjustments to the development of the systems’ particular parts ensuring their compatibility with your work style.

In many situations, an idea put forward to the developers will not be translated into the product. Why? An idea for a solution is just an idea for a solution – it may turn out that it doesn’t harmonise with the vision or product development strategy, or that there’s another way to respond to your business need. Therefore, you can be offered a solution different from what you initially had in mind, as at the end of the day, every problem can be solved in many ways. Don’t be afraid to try new available functions in practice, and if there is something you find unclear or if you aren’t sure whether you’re using the product’s full potential, be certain that when contacting its makers you will use all necessary guidance or assistance – after all, you live in the age of customers creating products!

The post Welcome to the era of customers creating products – why is your opinion so important? appeared first on Upsaily blog.

Machine learning mechanisms, analyzing thousands or millions of purchase transactions are searching for repetitive patterns. Such patterns present, for example, which customers are loyal to specific fashion brands, which products they need currently or what is the sequence of purchases.

If you want to learn how to apply machine learning in other e-commerce industries, check out this post!

The post Why is it worth using machine learning in the fashion industry? appeared first on Upsaily blog.

Product recommendation

This one is pretty simple and was already mentioned before. The outcome of this technique is a simple set of product/service rules based on customer product purchasing behavior. For example: if a customer bought milk, is he going to buy eggs too? In general: does buying one specific item increase the chances of buying other items? With a good dataset, ML algorithms can answer those questions and recommend the right products, consequently increasing the sale.

Customer segmentation

This tool divides all customers into statistical segments. They become groups of individuals that are similar in specific ways relevant to marketing, such as age, gender, interests and spending habits. It also helps companies determine which segments are the most or least profitable so they can adjust their marketing budgets accordingly.

Customer geolocation analysis

Analyzing and mapping information can allow companies to accurately predict customers’ habits. Marketing campaigns targeting specific consumer groups can be created using this data and pushed to consumers through their mobile device when a purchase is expected. Combining behavioral information with geotargeting can help companies develop cost-effective regional campaigns.

Customer lifetime value to customer acquisition cost (LTV/CAC)

On the visualization each customer is represented by one dot, while the colour refers to the segment’s number. The customers are devided by a recency of their last order, a frequency of the purchases, an average value of orders as well as their loyalty. The segments grup customers with similar behaviours. To maximize the effects of marketing, we should determine a segment specific way to communicate with the customers.

The ratio measures the relationship between the lifetime value of a customer and the cost of acquiring that customer. Knowing that ideal LTV/CAC ratio should be 3:1, companies use RFM (recency, frequency & monetary value) method offered by ML to calculate exact customer lifetime value and adjust the costs. In e-commerce, as in any other business, companies neither want to spend too much (unnecessary cash losses) nor too little (risk of losing the customer).

Proper price adjustments

Price optimization and price elasticity are two important aspects of every business. Relying on ML companies can create competitive and also contextually relevant prices.  Machine learning is being used today to determine pricing elasticity not only by each product, but also by client personal preferences, for example, if he or she prefers to buy on sale or not.

The above-mentioned steps are only a tip of an iceberg which machine learning is today. New tools are created every day, giving companies dozens of possibilities to use ML both in e-commerce oriented companies and other business branches.

The post 5 ways to use machine learning in e-commerce appeared first on Upsaily blog.